Measure Your Time

I’ve been thinking more and more about what the world looks like in an AI native, AI default, or AI first environment, and I personally can’t help but think about how we’ll end up spending a lot more of our time with AIs. There is an interesting chart floating around on LinkedIn that is arguing that social media usage peaked in 2022 (see below).

I like this chart because it’s clean and straightforward, it lets the data tell the story that social media consumption seems to have peaked, but still takes up a massive % of our time spent, and as a result of that time spent it’s incredibly valuable.

Where my mind goes to is, “well what does our time spent with AI look like today?” Is the megatrend of AI following a similar consumption curve as social media? From a revenue perspective the answer is yes. OpenAI is north of a $12bn run rate revenue, and Anthropic has crossed the $5bn run rate threshold, but what does it look like from a time spent perspective?

My guess is that many of us, despite being daily or monthly active users of products like ChatGPT and Claude, we actually probably OVERESTIMATE the time that we spend within these tools, and that’s an incredibly bullish signal on the opportunity ahead. My guess is the average person uses ChatGPT or another model provider, less than 30 minutes per day. The rough approximation would say that, based on social media consumption patterns, “time with AI” could increase 5x from here, but that’s before we account for last week’s idea of parallel intelligence. In order to get value out of AI, you don’t have to work or experience it sequentially, work can get done in parallel, which has never been possible for an individual person to do.

Parallel work has always taken human coordination and organization, and now it can be done via API calls in natural language.

What a world we live in.